摘要
Wind filed analysis from synthetic aperture radar image allows the estimation of wind direction based on image character.To make better use of wind streaks information from the calibrated SAR image, the wind direction is retrieved by the wavelet transform associated with spectral processing and two-dimensional Fast Fourier Transform.Basing on the good localization property of wavelet transform in time domain and frequency domain, we apply existing decimated wavelet transform approaches, and the purpose is to extract more reliable wind directional information.In the end, we compared our results with QSCAT/NCEP data and ship observation data, respectively.The results show that the Daubechies discrete wavelet transform can extract very good wind streaks information from SAR images, and the retrieval accuracy is better than the distance weighted gradient method.